Hadoop on Windows - "Error JAVA_HOME is incorrectly set."

Question

I'm trying to get Hadoop running 'Local Mode' on my Windows machine. I've been using this guide: http://hadoop.apache.org/docs/r1.2.1/single_node_setup.html

When I run the command: bin/hadoop, I get the following error message:

Error: JAVA_HOME is incorrectly set. Please update C:...\hadoop-2.7.1\conf\hadoop-env.cmd

Inside hadoop-env.cmd, I have the line: set JAVA_HOME=%JAVA_HOME%

When I type echo %JAVA_HOME% at the command line I get: C:\Java\jdk1.0.8_51

Even if I change the line in hadoop-env.cmd to: set JAVA_HOME=C:\Java\jdk1.0.8_51 I get the same error...

How could it be that my JAVA_HOME is incorrectly set?

Backstroke

Try setting JAVA_HOME environment variable at a system level not just your shell (assuming you just set it in the shell). If not then run 'Java -version' at a command prompt. – Eris

Thanks for your response Heath. I have set it at a system level. I have a system variable JAVA_HOME=C:\Java\jdk1.0.8_51 ; and I have C:\Java\jdk1.0.8_51\bin in my PATH user variable. Typing in Java -version at the prompt returns -> "java version 1.8.0_51..." – Backstroke

Does opening a new prompt and running hadoop work now – Eris

No. Sorry, I should have included a mention of that in my previous comment. – Backstroke

check this to resolve your issue https://mcmap.net/q/512853/-hadoop-installation-on-windows – Fencing

Answer 1

If your JAVA_HOME path contains spaces, you must use the Windows 8.3 Pathname

Augustin

is right! Thanks for saving my day. Just want to make it more detial, if you use the default setting while installing JAVA, you will have to set the JAVA_HOME to something like "C:\Progra~1\Java\jdk1.8.0_xxx". (Don't forget to replace xxx to your actual jdk update id.) – Forsta

Right! I also encountered the JAVA_HOME problem. But then, I got another error message "Error: Could not find or load main Trung" – Pappus

@Duong edit set HADOOP_IDENT_STRING=%USERNAME% to HADOOP_IDENT_STRING=username(without space) in hadoop-env.cmd file – Plovdiv

Answer 2

like the other answers, your java environment path must not contain space. The solution is as follows:

1. In the cmd line, charge the directory that contain the jdk (in my case C:\Program Files\Java\jdk1.8.0_73).

2. execute the following line:
   for %I in (.) do echo %~sI
   to display the short name of your installed jdk (in my case C:\PROGRA~1\Java\JDK18~1.0_7)

3. in the file "hadoop-env.cmd", change the line "JAVA_HOME=%JAVA_HOME%" with "JAVA_HOME=C:\PROGRA~1\Java\JDK18~1.0_7".

4. run again the file "hadoop-env.cmd" and it will work correctly.

Cordelia

That was awesome. Saved my day! Thank you a lot! – Predigestion

Answer 3

 set JAVA_HOME=C:\Program Files\Java\jdk1.8.0_131\

Above configuration was throwing same error as yours. I updated above with below in hadoop-env.cmd which is infact SFN

set JAVA_HOME=C:\Progra~1\Java\jdk1.8.0_131\

This allowed me to change/have different java version than that of system.

Shortlived

Answer 4

Try to place the java path in double-coats. Example:

JAVA_HOME="C:\Program Files\Java\jdk-12.0.1"   

It should work fine.

Skeens

That's not working. spark-shell for example will throw error. – Katushka

Answer 5

Use "%JAVA_HOME%" in conf\hadoop-env.cmd

Obed

Answer 6

if your java environment path contains space, such as "C:\Program Files\java\xxxxx" , the word 《Program Files》 contains a space, so CMD can't identificate

this is the right answer

Adeno

Answer 7

Try removing @ from @rem in hadoop-env.cmd and set JAVA_HOME=%JAVA_HOME%

rem The java implementation to use.  Required.

set JAVA_HOME=%JAVA_HOME%

Then try running hdfs namenode -format

This solution worked for me

Anthe

Answer 8

Explanation : C:\Program Files\Java\jdk-11 If you set this path to "System Environment Variables", Java will work normally on your machine. (jdk11 can be different in your case)

But when it comes to Hadoop, it looks like this in "hadoop-env.cmd" file :

set JAVA_HOME=C:\Program Files\Java\jdk-11 : "hadoop-env.cmd" file will ignore path after space : C:\Program + Files\Java\jdk-11 will be treated as two different things.

Solution :

step 1 : Copy "Java" folder from C:\Program Files\Java and paste it in "C:" directory, it will look like this : "C:\Java"

step 2 : Now edit "hadoop-env.cmd" file : set JAVA_HOME=C:\Java\jdk-11

step 3 : make sure to give same path in "System Environment Variables" :

"JAVA_HOME" : C:\Java\jdk-11 & "path" : C:\Java\jdk-11\bin

Restart your IDE.

• Alternate Solution : If you want to go with same path with spaces, then you should read about it here : using-spaces-in-windows-cmd-file-or-batch-file

Hope this helps.

Pairoar

Answer 9

In hadoop-env.cmd I changed %JAVA_HOME% to "C:\Program Files\openjdk-17_windows-x64_bin\jdk-17" with double quotation. It worked for me

set JAVA_HOME="C:\Program Files\openjdk-17_windows-x64_bin\jdk-17"

Brunner

Answer 10

in hadoop-config.xml， the logic is

if not exist %JAVA_HOME%\bin\java.exe ( echo Error: JAVA_HOME is incorrectly set. echo Please update %HADOOP_HOME%\conf\hadoop-env.cmd goto :eof )

if your java environment path contains space, such as "C:\Program Files\java\xxxxx" , the word 《Program Files》 contains a space, so CMD can't identificate

SO your can change the path instead.

Prefect